From: MITSUNARI Shigeo <[email protected]>
Changes since v8 (all in [1/2], addressing Jeff Law's review):
- Replace the goto with an if (quotient == NULL_RTX) block and
reindent the classic sub/shift/add sequence accordingly. The
skipped code itself is unchanged.
- Add gcc_checking_assert (post_shift <= size) and comments to
expand_wide_mulh_udiv documenting why the magic shift count is
never negative (choose_multiplier guarantees post_shift <=
ceil_log2 (d) <= size) and why the magic constant cannot overflow
(it is below 2^(2*size) and 2*size <= HOST_BITS_PER_WIDE_INT is
already checked).
- [2/2] is unchanged (approved by Uros Bizjak).
MITSUNARI Shigeo (2):
expmed: Optimize 32-bit unsigned division by constants on 64-bit
targets
i386: Add BMI2 MULX pattern for highpart-only multiplication
gcc/config/i386/i386.md | 15 +++
gcc/expmed.cc | 105 +++++++++++++++---
.../gcc.target/i386/bmi2-mulx-highpart-1.c | 23 ++++
gcc/testsuite/gcc.target/i386/mulq-highpart.c | 22 ++++
4 files changed, 149 insertions(+), 16 deletions(-)
create mode 100644 gcc/testsuite/gcc.target/i386/bmi2-mulx-highpart-1.c
create mode 100644 gcc/testsuite/gcc.target/i386/mulq-highpart.c
--
2.43.0
